Text summarization in NLP: From Extractive to Abstractive and Beyond
🎓 Text Summarization in NLP: From Extractive to Abstractive and Beyond
👩💻 Presented by: Sindhuja Lakkapally, Mahesh Babu Pallapolu, and Ashirwad Polumuri
📚 Course: CAP 6640-042 – Natural Language Processing
👨🏫 Instructor: Prof. Dingding Wang
🏫 Florida Atlantic University (FAU)
🔍 About the Presentation:
In this 15-minute group presentation, we explore how Natural Language Processing (NLP) enables machines to summarize large volumes of
text efficiently and intelligently. We walk through the evolution of text summarization — from extractive techniques that select key sentences, to abstractive methods that generate human-like summaries, and finally to large language model (LLM)-based approaches that can summarize across languages and media.
💡 Topics Covered:
✅ What is Text Summarization?
✅ Extractive vs. Abstractive vs. Hybrid Approaches
✅ Key Algorithms (TF-IDF, TextRank, Transformers)
✅ Evolution from Seq2Seq to Transformers (BERT, T5, BART, Pegasus)
✅ Applications in Media, Business, Healthcare, and Education
✅ Challenges – Faithfulness, Evaluation, Long Documents, Multilingual Issues
✅ Popular Datasets (CNN/DailyMail, XSum, Gigaword)
✅ Role of LLMs (GPT-4, Claude, Gemini)
✅ Future Trends – Multimodal, Ethical, and Personalized Summarization